Before any project
AI assessment
Use cases ranked by priority, and the Law 25 inventory of what would be sent to a model.
$3,000 to $5,000
Artificial intelligence
AI for Quebec SMBs
SignalOrange helps small and mid-sized businesses in Quebec use artificial intelligence where it saves time: automating repetitive work and informing a decision. We start with an assessment that inventories what would be sent to a model, then run a pilot on your own data, at a fixed price.
What we deliver
From one specific task to a tool used every day
A useful AI project starts with one specific task and data you can explain. We deliver ranked use cases, a pilot you can judge on facts, then an integration into your tools, while keeping control of what goes to a model.
Ranked use cases
We list the tasks where AI can help, estimate the effort and the gain for each, and rank them. If a task does not need AI, we tell you.
A Law 25 inventory
For each use case, we note which information would be sent to a model, to which provider and why. It is the starting point for the assessment Law 25 requires before information leaves Quebec.
A filter before anything leaves
The pilot runs through Router, our AI gateway. Its Guard removes personal information, secrets and tokens before the model is called. This filtering is a technical control and does not, on its own, ensure legal compliance.
Integration and training
In production, the AI is built into the tools your team already uses, and we train the people who will use it every day.
Readable data
Julie Choquette, our data analysis consultant, turns the data into decision dashboards. These dashboards show whether the use case produces the expected results.
Is it the right time?
An AI project makes sense if
- a repetitive task comes back every day or with every file;
- your team often looks for the same information in your documents;
- you want to know what goes to a model before you start.
Wait if
- the goal is to "do AI" without a specific task;
- nobody on your side will be able to judge the pilot's results;
- the data you would need does not exist yet.
Packages and prices
Four packages, from assessment to production
Each package has a fixed price within the published range. The assessment stands on its own: you can stop there and keep the list of use cases and the inventory.
4 to 6 weeks
Pilot
A pilot on your data, with Router filtering what leaves.
$10,000 to $20,000
After the pilot
Production rollout
Integration into your tools and training for your team.
$20,000 to $40,000
1 day
One-day workshop
A day of training for your team, with no project attached.
about $2,500
Fixed price, confirmed after a 30-minute call. In Canadian dollars, taxes extra.
Funding
Possible funding
Two public programs can support an AI project in a Quebec SMB. Eligibility is decided by the organization that runs each program: check with them before counting on any support.
| Program | Organization | What it covers | Official page |
|---|---|---|---|
| ESSOR, component 1B | Investissement Québec | A digital diagnostic, a digital plan and system selection, including AI integration where relevant. Businesses with 250 employees or fewer and revenue of at least $2.5M. Non-repayable contribution of up to 50% of eligible expenses, at most $20,000. The program ends on March 31, 2027. Official page in French. |
Read the official page
checked on 2026-11-06 |
| Regional Artificial Intelligence Initiative | Canada Economic Development for Quebec Regions | SMEs that are ready to adopt AI technologies or solutions. Repayable financial support of up to 50% of authorized costs. |
Read the official page
checked on 2026-11-06 |
How it works
How an engagement runs
-
Discovery
A 30-minute call, then the assessment. We look at the tasks, the data and the tools, and inventory what would go to a model.
-
Iterations
The pilot runs on your data, in small versions. We judge the results with your team before talking about production.
-
Handover
We integrate the solution into your tools, train the team and hand over the documentation, including the list of what is sent to which provider.
Who leads it
Jacob Roy, with Julie Choquette on data
Jacob Roy leads every AI engagement. Julie Choquette, our data analysis consultant, prepares the data and the dashboards used to judge the pilot.
Meet the whole team
Jacob Roy
Founder and Solutions Architect
Leads the engagement and the integration.
With ten years in solution architecture and software development, Jacob builds things that last and that fit the way you work.
LinkedIn
Julie Choquette
Consultant, data analysis
Data and decision dashboards.
Julie turns complex data into decision dashboards and concrete actions, to help you interpret your numbers.
LinkedInWhat we have built
AI in our own products
We use AI in the software we design and run. A client's pilot goes through the same gateway as our products.
Router
Our OpenAI-compatible AI gateway: more than 40 models behind one API, automatic failover, Guard filtering sensitive data before it is sent, logs hosted in Montreal and a bill in Canadian dollars.
See the page
Meridian
Canadian CRM software with an AI assistant on every screen, with nothing to configure. Sensitive data is filtered before anything is sent to an external model.
See the page
Estrade
A hockey pool as a progressive web app, with an MCP server that lets an AI agent read the standings, the team and the schedule.
See the pageBefore you send data
Law 25 and AI APIs
Our guide explains what a Quebec SMB can check before sending personal information to an AI API: an inventory of prompts, the assessment before information leaves Quebec, and the questions to ask the provider. It is not legal advice.
A task AI could take off your plate?
In 30 minutes, we look at the task and the data it touches. We will tell you which package fits and at what fixed price, or whether AI is not the right answer.